Contribution of polyfluoroalkyl phosphate esters (PAPs) and other precursor compounds to perfluoroalkyl carboxylates (PFCAs) in humans and the environment

(1)

Contribution of polyfluoroalkyl phosphate esters (PAPs) and other precursor compounds to perfluoroalkyl carboxylates (PFCAs) in humans

and the environment

(2)

If, having endured much, we have at last asserted out "right to know,"

and if by knowing, we have concluded that we are being asked to take senseless and frightening risks, then we should no longer accept the coun- sel of those who tell us that we must fill our world with poisonous chemi-

cals; we should look about and see what other course is open to us.

Rachel Carson, Silent Spring

(3)

Örebro Studies in Chemistry 18

ULRIKA ERIKSSON

Contribution of polyfluoroalkyl phosphate esters (PAPs) and other precursor compounds to perfluoroalkyl carboxylates (PFCAs) in humans and the environment

(4)

Title: Contribution of polyfluoroalkyl phosphate esters (PAPs) and other precursor compounds to perfluoroalkyl carboxylates (PFCAs) in humans and the

environment

Publisher: Örebro University 2016 www.oru.se/publikationer-avhandlingar

Print: Örebro University, Repro 10/2016 ISSN1651-4270

ISBN978-91-7529-164-2

(5)

Abstract

Ulrika Eriksson (2016): Contribution of polyfluoroalkyl phosphate esters (PAPs) and other precursor compounds to perfluoroalkyl carboxylates (PFCAs) in humans and the environment. Örebro Studies in Chemistry 18.

Per-and polyfluoroalkyl substances (PFAS) are anthropogenic compounds that have been spread all over the world. The use of fluorotelomer compounds, short-chained homologues, and other PFASs with perfluorinated moieties has emerged recent years. One of these emerging compound classes is polyfluoroalkyl phosphate esters (PAPs), which have the ability to degrade into persistent PFCAs.

The aim of this thesis was to assess the contribution of PAPs and other precursors to the exposure of PFCAs to humans and the environment.

The main objective was to analyze a wide range of PFAS in human serum, wild bird eggs, indoor dust, waste water, and sludge.

There was a significant contribution from selected precursors to the total amount of PFASs in the abiotic compartments indoor dust, waste water, and sludge. Levels of PAPs found in house dust exceeded those of PFCAs and perfluorosulfonic acids (PFSAs), revealing PAPs as a world-wide important exposure source.

A net increase was during waste water treatment was observed for several PFASs in Swedish waste water treatment plants. Together with presence of precursor compounds and intermediates in the influent water and the sludge, this suggest that degradation of PFCA precursors contributed to the increase of PFCAs. Detection of precursors in human serum, together with slow declining trends of PFCAs, revealed an ongoing exposure of PFCAs to the general population of Australia.

The diPAPs and the FTSAs were also detected in raptor bird eggs from Sweden from both the terrestrial and the freshwater environment. The precursors concentrations and patterns observed reveal that current regulatory measures are insufficient for the purpose of protecting humans and the environment from PFASs exposure.

Keywords: PAPs, precursors, PFCA, exposure, indoor dust, human serum, WWTP, bird eggs

Ulrika Eriksson, School of Science and Technology, Örebro University, SE-701-82 Örebro, Sweden, ulrika.eriksson@oru.se

Abstract

(6)

(7)

List of papers

This thesis is based on following papers:

Paper I:

Eriksson U. Kärrman A. 2015. World-wide indoor exposure to polyfluoroalkyl phosphate esters (PAPs) and other PFASs in household dust. Environmental Science and Technology 49 (24), 14503-14511. Re- produced with permission from Environmental Science and Technology.

Paper II:

Eriksson U, Mueller J, Toms L-M L, Hobson P, Kärrman A. 2016. Tem- poral trends of PFSAs, PFCAs and selected precursors in Australian serum from 2002 to 2013. Accepted for publication in Environmental Pollution.

Paper III:

Eriksson U, Haglund P, Kärrman A. 2016. Per- and polyfluoroalkyl substances (PFASs) in sludge and water from Swedish waste water treatment plants (WWTP). Under review in Journal of Environmental Science.

Paper IV:

Eriksson U, Roos A, Lind Y, Hope K, Ekblad A, Kärrman A. 2016. Com- parison of PFASs contamination in the freshwater and terrestrial environments by analysis of eggs from osprey (Pandion haliaetus), tawny owl (Strix aluco), and common kestrel (Falco tinnunculus). Environmental Research 149, 40-47.

(8)

Papers not included in this thesis:

Eriksson U, Kärrman A, Rotander A, Mikkelsen B, Dam M. 2013. Per- fluoroalkyl substances (PFASs) in food and water from Faroe Islands.

Environmental Science and Pollution Research 20 (11) 7940-8.

Ekblad A, Mikusinska A, Ågren GI, Menichetti L, Wallander H, Vilgalys R, Bahr A, Eriksson U. 2016. Production and turnover of ectomycorrhizal extramatrical mycelial biomass and necromass under elevated CO2 and nitrogen fertilization. New Phytologist 211(3) 874-885.

Geng D, Ericson Jogsten I, Kukucka P, Eriksson U, Ekblad A, Grahn H, Roos A. 2016. Temporal Trends of Polychlorinated Biphenyls, Organo- chlorine Pesticides and Polybrominated Diphenyl Ethers in Osprey Eggs in Sweden over the Years 1966 – 2013. Submitted to Environmental Pollu- tion.

(9)

List of abbreviations

Name Abbreviation

AFFF Aqueous film forming foam

diPAP Polyfluoroalkyl phosphoric acid diester ECF Electrochemical fluorination

EtFOSA Ethyl perfluorooctane sulfonamide EtFOSE Ethyl perfluorooctane sulfonamidoethanol FBSA Perfluorobutane sulfonamide

FTAC Fluorotelomer acrylate FTAL Fluorotelomer aldehydes FTCA Fluorotelomer carboxylic acids

FTMAP Fluorotelomer mercaptoalkyl phosphate diester FTOH Fluorotelomer alcohol

FTSA Fluorotelomer sulfonic acid

FTUCA Fluorotelomer unsaturated carboxylic acids MeFBSA Methyl perfluorobutane sulfonamide MeFOSA Methyl perfluorooctane sulfonamide MeFOSE Methyl perfluorooctane sulfonamidoethanol monoPAP Polyfluoroalkyl phosphoric acid monoester PAP Polyfluoroalkyl phosphate ester

PBSF Perfluorobutane sulfonyl fluoride PFAS Per- and polyfluoroalkyl substance PFBA Perfluorobutanoic acid

PFBS Perfluorobutane sulfonic acid PFCA Perfluoroalkyl carboxylic acid PFDA Perfluorodecanoic acid

PFDPA Perfluorodecyl phosphonic acid PFDS Perfluorodecasulfonic acid PFEI Pentafluoroethyl iodide PFHpA Perfluoropentanoic acid PFHpS Perfluoroheptasulfonic acid PFHxA Perfluorohexanoic acid PFHxA Perfluorohexadecanoic acid PFHxPA Perfluorohexyl phosphonic acid PFHxS Perflurohexasulfonic acid PFNA Perfluorononaoic acid PFNS Perfluorononasulfonic acid PFOA Perfluorooctanic acid

(10)

PFOcDA Perfluorooctadecanoic acid PFOPA Perfluorooctyl phosphonic acid PFOS Perfluorooctasulfonic acid PFPA Perfluorophosphonic acid PFPA Perfluoroalkyl phosphonic acid PFPeA Perfluoropentanoic acid PFPeS Perfluoropentane sulfonic acid PFPiA Perfluorophosphinic acid PFPiA Perfluoroalkyl phosphinic acid PFSA Perfluoroalkyl sulfonic acid PFTDA Perfluorotetradecanoic acid PFTrDA Perfluorotridecanoic acid POSF Perfluorooctane sulfonyl fluoride PTFE Polytetrafluoroethylene

PVDF Polyvinylidene fluoride

SAmPAP Perfluorooctane sulfonamide phosphate esters TDI Tolerable daily intake

TFE Tetrafluoroethylene

triPAP Polyfluoroalkyl phosphoric acid triester WWTP Waste water treatment plant

(11)

1. Introduction

The group of per- and polyfluoroalkyl substances (PFASs) consist of a large number of classes, of which most of them have in common that they have an alkyl chain which is partially or fully fluorinated, with a function- al head group, typically carboxylate, sulfonate, phosphate, or alcohol (Buck et al. 2011). The high electronegativity of the fluorine provides strong polarity and high energy to the carbon-fluorine-bond (Chambers 2009). The bond strength increase with number of fluorines attached to the carbon atom. The strong carbon-fluorine bond together with effective shielding of the alkyl chain of the fluorine atoms result in high stability of the molecule. Fluorine has a high ionization potential and therefore the inter- and intramolecular interactions in fluorocarbons are low, leading to extremely low surface tension and similar or higher volatility compared to their hydrogen carbon counterparts. A charged moiety attached to the fluorinated alkyl chain enhance the water solubility, resulting in a molecule with both hydrophobic and hydrophilic properties. These superior properties compared to other surfactants make PFASs suitable for oil- and water repellency and high temperature applications, which have led to an extensive use in a wide range of applications such as in wetting and leveling agents, paints, coatings, waxes, chrome plating bath, fire-fighting foams, cosmetics, paper, food packaging, textiles, carpets, cleaning agents, pesticides, photographic emulsifiers. The PFAS group comprises several thousand compounds (KEMI 2015). Only a few of these are usually covered by monitoring and scientific studies. Analysis of total extractable organic fluorine content suggest that a large proportion of the organofluo- rine in the environment and humans are unknown compounds (Yeung et al. 2013a, Yeung and Mabury 2016). Commonly used commercial PFASs that will eventually be transformed to persistent PFASs, but are always ignored in monitoring programs, may represent a significant portion of the total amount.

1. 1. Synthesis and use of PFASs

There are two main processes for PFAS production; electrochemical fluorination (ECF) and telomerization. In the ECF process, the starting material is octane sulfonyl fluoride (C8H17SO2F) or a carbonyl fluoride (C7H15COF) (3M Company 1999a, Lehmler 2005, Buck et al. 2011). In the presence of anhydrous fluoride, a current is passed through the solution and all the hydrogen atoms are replaced with fluorine, leading to a perfluorinated

(14)

substance (fig. 1). This process can lead to rearrangement and breakage of the carbon chain, yielding a proportion of 20-30% branched isomers.

With octane sulfonyl fluoride as starting material, the process yields perfluorooctane sulfonyl fluoride (POSF), which can be further processed to yield perfluorooctane sulfonic acid (PFOS) through base-catalyzed hydrol- ysis, and with carbonyl fluoride the process yields perfluorooctanoic acid (PFOA). PFOS has been produced using ECF since 1949 and has been widely applied in for example aqueous film forming foam (AFFF), metal plating, and hydraulic fluids (Paul et al. 2009). Further reaction with methyl or ethyl amine yields perfluorooctane sulfonamides (MeFOSA and EtFOSA). FOSA can either be used in commercial applications, for example EtFOSA in pesticides (Gilljam et al. 2016), or be further reacted with ethylene carbonate to perfluoroalkane sulfonamidoethanols (MeFOSE and EtFOSE). MeFOSE has been used in polymeric substances in textiles and carpets (Olsen et al. 2005). EtFOSE was amongst others used as building material for perfluorooctane sulfonamide phosphate esters (SAmPAPs), which has been used for paper treatment (D'eon et al. 2009). POSF-based chemistry has been phased out by the former major manufacturers but homologues with shorter chain length, foremost perfluorobutane sulfonyl fluoride (PBSF), have replaced POSF in the ECF process (OECD 2007, Buck et al. 2011).

(15)

Figure 1. A simplified scheme over the ECF process (Buck et al. 2011)

In the telomerization process, tetrafluoroethylene (TFE) is reacted with a pentafluoroethyl iodide (PFEI) yielding perfluoroalkyl iodides (PFAI) (fig.

2) (Lehmler 2005, Buck et al. 2011). Almost exclusive linear homologues are produced, in contrast to the ECF process, though synthesis of branched perfluoroalkyl iodides also has been described (Bertocchio et al.

1993). In general, only even-numbered homologues are produced, though there have been some reports about odd-numbered fluorotelomer-based compounds in literature (Ding et al. 2012). The PFAI can be reacted to yield for example PFOA, perfluorononanoic acid (PFNA), fluorotelomer iodide (FTI), and perfluoroalkyl phosphonic acids (PFPA) and phosphinic acids (PFPiA) (Wang et al. 2016a). The major use of PFOA is as processing aid in the synthesis of the fluoropolymer polytetrafluoroethylene (PTFE), and for PFNA as a processing aid in synthesis of polyvinylidene fluoride (PVDF). Mixtures of PFPAs and PFPiAs have been used in pesticides, wetting and leveling agents, plating, and cleaning products (Dookhith 2001, Pilot Chemical 2016).

(16)

FTI can be further reacted to produce various PFASs, fluorotelomer alco- hols (FTOH), fluorotelomer olefins (FTO), fluorotelomer sulfonic acids (FTSAs), polyfluoroalkyl phosphate esters (PAPs) and perfluoroalkyl carboxylates (PFCAs). FTSA-based compounds have been used mainly in fire fighting foams as an active ingredient, due to their low surface tension, which enables aqueous film formation on hydrocarbon fuels (Harding- Marjanovic et al. 2015). Other uses have been described as well, for instance in ink to reduce puddling of the ink-jet ink on the nozzle plate (Ma et al. 2002). The FTOHs can be functionalized to fluorotelomer acrylate (FTAC) and further synthesized to fluorotelomer acrylate polymers, for which major use has been in textile, leather, and paper. Alternatively, the FTOHs can be further functionalized to yield PAPs or fluorotelomer mercaptoalkyl phosphate diesters (FTMAPs). Major use for PAPs have been in paper and packaging, including food packaging, but other uses have been described as well, as in personal care products, floor finish, cleaning products, paints, and coatings (Pilot Chemical 2016). The products in the telomerization process are usually mixtures that can have chains of a length between 2 and 18 fluorinated carbons. Distillation is used for separations and purifications. Neither ECF nor telomerization have a 100% yield, and the starting materials can be left as residuals in the final product.

(17)

Figure 2. A simplified scheme of the telomerization process (Lehmler 2005, Buck et al. 2011, Wang et al. 2016a).

1. 2. A historical view

PFASs have been produced since the 1950s. Amongst the first applications was the use of ammonium perfluorooctanoate (APFO), a salt of PFOA, as processing aid for polymerization of PTFE (Prevedouros et al. 2006). The fluorinated tail of APFO keep the PTFE particles in dispersion state during the polymerization process (McKeen 2015). Similarly, PFNA was implemented as processing aid for polyvinylidene fluoride (PVDF). Between 1951 and 2002, ECF was the dominating process for manufacturing both PFOS and PFOA (Prevedourus, Buck). The 3M company was the major global POSF producer with manufacturing plants in the US and Europe, though minor quantities were also produced by other companies in Eu- rope, Asia, and South America (Paul et al. 2009). Fluorotelomer production started in 1961 and PFOA production using the telomerization pro-

(18)

cess started in the 1970s by DuPont (Wang 2014), though ECF continued for decades to be the major PFOA synthesizing technique (Prevedouros et al. 2006, Wang et al. 2014). In the 1970s, elevated levels of organic fluorine were found in blood of workers at a PFAS production factory, and PFOA was found in their urine (Ubel et al. 1980). This was the first indi- cator of human exposure to this compound group.

From 1960s to 2000, production volumes of PFASs steadily increased.

Annual APFO production has been estimated to increase from 5 – 25 ton in the early 1960s to 200 – 300 ton in the late 1990s (Prevedouros et al.

2006). Production of PFOSF has been estimated to increase from about 500 ton in 1970 to 4 650 ton in 2000 (Paul et al. 2009). Of this amount, 33% was used in paper and packaging, including food packaging, 48% in surface treatment such as textile and leather protection, 3% in AFFF, and 15% in other consumer and industrial applications (3M Company 2000).

In the late 1990s, the analytical technique of electrospray ionization mass spectrometry (ESI-MS) had been largely improved and was applied for the analysis of PFASs, lowering the limit of detection for PFOS to 50 ppb (3M Company 1999b). Attention started to grow around PFAS and in particu- lar PFOS. In 1997, PFOS in serum of the general population was observed by the 3M Company (3M Company 1999b). Previously PFOS and PFOA had been found in occupational exposed workers and monitored for years, but now PFOS was also found in a control group with supposedly con- taminant-free serum. Between 1997 and 2000, the 3M Company reduced their waste water discharged by 50% and the air emission by 40% at their facility in Decatur, Alabama, one of their two major POSF plants (3M Company 2000). After increasing pressure from the US EPA, the 3M Company announced their phase-out of the C8-technology in May 2000 and completed it in 2002 (3M Company 2000). This led to several changes in the PFAS production and manufacturing.

As a result of ceased production in North America and Europe, PFOS production expanded in China. The annual production of PFOS in China has increased from <50 ton in 2001 to 250 ton in 2006 and ranged between 220 and 250 ton in 2006 - 2011 (Zhang et al. 2012, Xie et al.

2013). The 3M Company continued their PFAS production, but switched over to PBSF-based chemistry (3M Company 2002). New PFASs with sulfonic acids have also been introduced worldwide, for example perfluoroalkyl ether sulfonic acids (PFESAs) (Wang et al. 2013).

(19)

In the years 2006 – 2009, the use of PFOS became further restricted. In the EU, the use and import of PFOS was restricted under REACH in 2006 (REACH 2010). Norway banned the use of PFOS in AFFFs, impregna- tions, and textiles in 2004 (Norway 2004). Manufacturing, use and sale become prohibited in Canada in 2008 (Canada 2006). PFOS was listed into the Annex B of Stockholm Convention on Persistent Organic Pollu- tants in 2009 (UNEP 2009). Recently, a project has been initiated by For- eign Economic Cooperation Office (FECO) and World Bank (WB) with the aim to help China reduce their PFOS emission (GEF 2016).

Restrictions of PFCAs have been lagging and implemented later than for PFOS. The 3M Company produced 85% or more of the global production volume of PFOA until 2000, using the ECF technique (OECD 2002).

DuPont had previously purchased PFOA from the 3M company, but started producing PFOA through fluorotelomerization at their facility in the US in 2002 (OECD 2007).

In 2006, the Voluntary Stewardship Program was initiated by the US EPA, where eight major PFAS manufacturers participated (DuPont, Arkema, Asahi, BASF Corporation, Clariant, Daikin, 3M/Dyneon, and Solvay So- lexis), aiming to reduce PFOA emissions by 95% to 2010 and completely phase-out by 2015 (EPA 2006). Fluorotelomer companies not bound to the Stewardship Program, mainly in China, India, and Russia, have taken over their market share of long-chain PFCAs (Wang et al. 2014). It has been estimated that 31% of the fluorotelomer production was covered by companies not participating in the Stewardship Program in 2013, of which 74% of the production occurred in China (ECHA 2014). The process of regulation of PFOA and long-chain PFCAs in the EU began in 2012-2013 by adding these compounds to the REACH Candidate List of Substances of Very High Concern (SVHC) (REACH EC No. 1907/2006). In 2012, the production volume of PFOA in China was 90 ton, exclusively produced by the ECF process (Li et al. 2015). Emissions of PFOA at one production site in China have been estimated to be 58 ton in 2013 (Wang et al.

2016b). In comparison, 50 ton PFOA was emitted in 2006 from the eight companies participating in the Stewardship program.

A majority, about 80%, of the fluorotelomers were incorporated in polymers in year 2000, while the remaining 20% were used in surfactants (TRP 2002). The information about actual production volumes of

(20)

fluorotelomer compounds are however scarce. It has been estimated though that the annual global production was 2 500 ton/year in 1961 – 1979, then increased to 7 500 ton/year in 1980 – 1994, followed by an increase to 20 000 ton/year in 1995 – 2004 (Wang et al. 2014). Current production volume (2004 – 2030) is estimated to be 45 000 ton/year.

Majority of replacement compounds for the long-chained PFASs that have entered the market continues to have a fluorinated structure. One of the major changes is the shift in the industry towards short-chain C4 – C6 PFASs (3M Company 2002, Daikin 2007, DuPont 2008, AGC 2016). The shift towards short-chain PFASs also includes PAPs. For example, the company Daito Kasei has replaced their product PF, a mixture of C8 – C20 mono-, di-, and triPAPs, with 6:2 monoPAP under the trade name EPF (environmental PF) (Kasei 2015). Recently, the use of C8-C18 PAPs in food contact paper and paperboard was banned in the US (FDA 2016).

Other replacement compounds are the polyfluorinated ethers (PFPE), for instance ADONA (CF3OCF2CF2CF2OCHFCF2COO⁻NH4+) and GenX (CF3CF2CF2OCF(CF3)COO^-NH4+) from Dyneon and Chemours (formerly DuPont), respectively (Wang et al. 2013). New compounds are constantly entering the market, and it has been estimated that more than 3000 commercial PFAS compounds are circulating on the global market (KEMI 2015). Volumes of PFASs produced are also expected to increase as a consequence of short-chain PFASs having less effective surfactants properties compared to long-chain PFASs. Forecasts predict that the fluorotelomer industry will continue to increase as a consequence of increased living standard and high market demand on products such as textiles, papers, metal plating, and semiconductors (Insights 2016). There are no signs of an instant elimination of PFASs at a global scale within a perspicuous future.

1. 3. Direct and indirect sources

The exposure sources of PFCAs and PFSAs are commonly referred to as being of direct or indirect origins. Direct sources are herein described as including both intentionally produced compounds, residuals from the production process, and unintended produced byproducts. Indirect sources are herein described as compounds formed during degradation from precursor compounds. Due to limitations of information on actual production volumes and homologue distribution, and uncertainties in the yield of persistent PFASs from precursors, it has been difficult to determine to

(21)

which extent indirect sources contribute to total PFCAs and PFSAs emission. Prevedouros et al. (2006) made the first extensive estimation and concluded that direct emission was the major source for global PFOA and PFOS contamination.

More recently, a global emission inventory showed that emissions of PFOA, PFNA, PFUnDA, and PFTrDA historically (1951 - 2002) came from direct sources, while the majority of the short-chain PFCAs (C4 – C7) were originated from degradation of precursor compounds. Between 2003 and 2015, it was estimated that degradation of precursor compounds and impurities in fluorotelomer-based products were the dominant emission sources for most PFCAs (Wang et al. 2014).

Another possible indirect source that has been heavily debated is the potential contribution from degradation of fluorotelomer polymers (Wash- ington and Jenkins 2015). Their relatively long half-times observed in soil- plant microcosm induce a delay in PFCA release, and bring a possible future dramatic increase of persistent PFAS (Rankin et al. 2014).

1. 4. Degradation

Precursor compounds not having fully fluorinated alkyl chains may have the ability to degrade, in both biotic and abiotic mechanisms, to persistent PFAS. Various pathways for different PFASs in different environmental compartments has been suggested. Biodegradation of PAPs has been demonstrated in rats, sludge, and soil (D'Eon and Mabury 2007, Lee et al.

2010, Lee et al. 2014). It was shown that PAPs degraded to corresponding PFCAs, for example 8:2 diPAP degraded into PFOA.

Biodegradation of n:2 diPAP starts with cleavage of the phosphate ester bond resulting in production of n:2 monoPAP and n:2 FTOH (Lee et al.

2010) (fig. 3). Further cleavage of the phosphate ester bond of n:2 monoPAP will result in n:2 FTOH. Subsequent degradation then follows the FTOH degradation pathway, which has been thoroughly studied in soil, sludge, and microbial culture (Dinglasan et al. 2004, Liu et al. 2010, Kim et al. 2014a). The n:2 FTOH is oxidized to a n:2 fluorotelomer aldehyde, which is further oxidized to n:2 fluorotelomer saturated carboxylic acid (FTCA). Then n:2 fluorotelomer unsaturated carboxylic acid (FTUCA) is formed through microbial defluorination. Degradation then follows two different pathways. Further defluorination of n:2 FTUCA

(22)

yield either n-1:2 sFTOH or n-1:3 FTUCA. The terminal product of n-1:2 sFTOH will be Cn PFCA. In the other pathway, Cn-2 PFCA and Cn-1 PFCA will be the degradation product from n-1:3 FTUCA. There are differences in the FTOH degradation pathway depending on the chain length. Similar molar yield of PFHxA and PFPeA from 6:2 FTOH has been observed in activated sludge, while in degradation of 8:2 FTOH in soil, PFOA was the major degradation product with only a small fraction of PFHpA (Liu et al.

2007, Wang et al. 2011).

In mammals as rat and mice, and in fish, degradation pathways have simi- larities with microbial degradation. The only difference is that 8:2 FTOH has shown to also produce PFNA through α-oxidation of 8:2 FTCA in rats, mice, and fish (Martin et al. 2005, Fasano et al. 2006, Butt et al.

2010). Degradation of 8:2 FTOH to PFCAs has been suggested to occur in humans, with PFOA as major metabolite (Nilsson et al. 2013).

Proposed biodegradation pathway of 6:2 FTSA involves desulfonation as an initial step (fig. 3), and a subsequent pathway much in similar with FTOH degradation pathway (Wang et al. 2011). In the degradation pathway suggested by Wang et al., 1-hydroxy 6:2 FTS is formed after desulfonation of 6:2 FTS, followed by rapid conversion to 6:2 fluorotelomer aldehyde, bypassing the formation of 6:2 FTOH. 6:2 FTSA in activated sludge has shown to result in formation of PFPeA and PFHxA. The FTSAs can also be formed from degradation of 6:2 fluorotelomermercaptoalkyl- amido sulfonate (FTSAS), an ingredient in AFFF, which has been demonstrated in sludge biodegradation experiments (Weiner et al. 2013). The proposed pathway involves S-dealkylation of 6:2 FTSAS by P-450 to 6:2 fluorotelomer thiol, which can be further oxidized to 6:2 FTSA. Alterna- tively, oxygenation of 6:2 FTSAS yields 6:2 FTSAS sulfoxide, and subse- quently cleavage of the C-S bond finally yields 6:2 FTOH, which can be further degraded into PFCAs.

PFPiAs have shown to degrade to corresponding PFPAs, and PFPA/PFPiAs have also been suggested to be potential PFCA precursors (Lee et al.

2012). However, in a study of PFPA/PFPiA biotransformation in rainbow trout, no PFCAs above LOD were found (Lee et al. 2012).

In atmospheric degradation of n:2 FTOH, oxidation of n:2 FTOH leads to formation of n:2 FTAL, which will be further degraded to equal amounts

(23)

of Cn PFCA and Cn+1 PFCA and to a lesser extent homologues other than Cn PFCA (Ellis et al. 2004). The proportional yield thus differs between biotic and atmospheric degradation, and it has been hypothesized that an odd-even pattern in remote places is a result of atmospheric degradation, followed by increased bioaccumulation with increasing chain length (Mar- tin et al. 2004, Bossi et al. 2005).

(24)

Figure 3. Combined simplified biodegradation pathways for 6:2 diPAP, 6:2 monoPAP, 6:2 FTSA, and 6:2 FTOH (Dinglasan et al. 2004, Lee et al. 2010, Liu et al. 2010, Wang et al. 2011, Kim et al. 2014a)

1. 5. Exposure pathways and sources

PFASs are spread in the environment through point sources and diffuse sources. Point sources are production facilities, industrial facilities that utilize PFASs in their production, domestic waste water treatment plants (WWTPs), landfills, military sites, airports, and firefighting practicing

(25)

grounds. Recently, the magnitude of PFAS pollution from such point sources was highlighted in a comprehensive study from the US, where harmful levels of PFOA in drinking water were associated with emissions from industrial sites, military firefighting training areas, airports, and WWTPs, affecting 6 million people (Hu et al. 2016).

A number of studies have reported high levels of PFASs in fish, water, waste water, soil, and sediment in surroundings of manufacturing and production sites (Wang et al. 2010, Bao et al. 2011, Oliaei et al. 2013).

For example, groundwater levels of up to 82 µg/L PFOA and 31 µg/L PFOS has been found at a production site in the US (Oliaei et al. 2013), and PFOA levels of up to 48 ng/g and 668 ng/L has been found in river sediment and water close to a fluorochemical plant in China (Bao et al.

2011).

Fire fighting practicing areas are a significant point source, when PFAS- containing AFFF is used as extinguishing. The use of AFFF at fire fighting practicing areas has led to severe contamination of soil and groundwater worldwide (Gewurtz et al. 2014, Houtz et al. 2016). Elevated levels of PFOS, 6:2 FTSA, PFPeA, and PFHxA in waste water have been linked to AFFF sources connected to the WWTPs (Houtz et al. 2016). Studies of former fire fighting practicing areas have revealed that PFASs continue to be released to the surroundings for a long time, even decades (Ahrens et al.

2015, Arias et al. 2015, Filipovic et al. 2015)

Waste water treatment plants are a source for PFASs to water, sludge, and the atmosphere. Increased levels of PFASs in the effluent compared to the influent have been attributed to degradation from precursor compounds, such as PAPs, FTSAs, FTOHs, FOSAs, and FOSEs (Schultz et al. 2006). In the WWTPs, PFASs will be distributed to different environmental compartments depending on their physico-chemical properties. The more water-soluble short-chain PFASs will be distributed mainly to the water phase, while long-chain PFASs are more prone to sorb to the sludge. Vola- tile PFASs are emitted to the atmosphere, and FTOHs has been recognized as the dominant PFAS class released to the atmosphere from waste water treatment facilities (Ahrens et al. 2011a). However, a recent study reported similar atmospheric emission of PFCA/PFSAs as for FTOHs (Yao et al.

2016). Most WWTPs are not designed for efficient removal of PFASs.

Adsorption to activated carbon, nanofiltration membranes, and advanced oxidation processes have shown to be effective for removal of PFOS and

(26)

PFOA (Arvaniti and Stasinakis 2015). Other examples of processes that have been evaluated are reverse osmosis and reduction processes. Howev- er, short-chain PFASs are not effectively captured by these techniques.

While the aquatic environment is affected by the release of waste water, sludge will end up in landfills or is used as bio-solids. Application of sludge in agriculture contributes to further spread of PFASs through plant uptake (Yoo et al. 2011, Lee et al. 2014). Upon disposal of PFASs in landfills, there is a risk of leaching to the environment through migration in soil, reaching the groundwater, and emissions to the atmosphere (Ahrens et al. 2011a).

Except from point source emissions, humans and environment are subject- ed to diffuse exposure, which are usually related to as inputs from mainly dry- and wet deposition, oceanic currents, but also urban runoff from streets (Ahrens 2011).

The contribution from urban runoff is less studied compared to the release from WWTPs, but emissions have been estimated to contribute equally as WWTPs to PFAS mass load in urban rivers (Zushi and Masunaga 2009).

Currently, there are two hypotheses regarding the transport of PFASs;

oceanic transport and atmospheric transport. Oceanic transport is a slow process, where it can take years for pollutants to be transported to remote regions (Armitage et al. 2006). Volatile precursors such as FTOHs and FASAs, can be atmospheric transported long distances and have an atmospheric lifetime of several weeks before degradation (Ellis et al. 2003, Martin et al. 2006). The relative importance between oceanic and atmospheric transportati of PFAS to remote areas has been estimated, and PFOA and PFNA have been suggested to be predominantly oceanic transported, while atmospheric transport and precursor degradation were found to be of more significance for PFDA, PFUnDA, PFDoDA and PFTrDA (Armitage et al. 2009). In a global survey of soil from areas with no evi- dent human impact, the homologue pattern of PFASs indicated significant contribution from atmospheric degradation of precursors, based on the ratio of PFOA/PFNA, supporting the hypothesis that atmospheric transport is the dominant pathway for the rural and terrestrial environment (Rankin et al. 2016).

Atmospheric transported PFASs have been found to be a significant source for contamination in the aquatic environment, with similar or higher con-

(27)

tribution compared to emissions from waste water (Muller et al. 2011, Kim et al. 2014b). The relative importance of diffuse emissions compared to WWTPs have shown to be related to population density, even at sites with no known PFAS industry activity, and diffuse emissions became more important in less densely populated areas.

In wild-life, animals are subject to PFAS exposure from food and water through point-source emissions, and long-range atmospheric and oceanic transport. Higher levels of PFAS in the terrestrial environment compared to agrarian and close to conurbation environments have been observed in roe deer, suggestions diffuse atmospheric sources to be the dominant exposure pathway for terrestrial mammals (Falk et al. 2012).

Food, drinking water, and indoor environment have been identified as the major exposure pathways to PFASs for the general population (Vestergren and Cousins 2009, Domingo 2012). Of these sources, food has been re- garded to be the major contributor, especially fish, seafood and meat (D'Hollander et al. 2010, Domingo 2012, EFSA 2012). As mentioned previously, food items such as vegetables can be contaminated by PFASs through soil and water during growth. Plants have shown to take up PFASs from soil in greenhouse and field experiments, especially short- chain PFASs (Lee et al. 2014). Animals used for food production take up PFASs through ingestion of contaminated food and water, and many PFASs bioaccumulate along the food chain. Contamination can also occur through migration from food packaging material (Begley et al. 2005).

Drinking water is another important exposure pathway. Elevated PFAS concentrations in humans have been linked to contaminated point sources such as firefighting practicing grounds, but there are also reports suggest- ing that even low water PFAS concentrations could implicate a risk to human health (Post et al. 2012, Weiss et al. 2012, Hu et al. 2016).

Indoor environment has been recognized to be a significant route of exposure (Bjorklund et al. 2009). Route of exposure is mainly through dust for ionic PFASs and through air for neutral, volatile PFASs. The uptake from dust is believed to occur mainly through ingestion, and the dermal uptake is only a small fraction of the total dust intake (Lorber and Egeghy 2011).

(28)

Although exposure assessments suggested rather strong positive correla- tion between PFOS levels in serum and diet, the association between PFOA in serum and food intake was found to be weaker, implying other important exposure pathways for PFOA (Fromme et al. 2007, De Felip et al. 2015). Exposure assessments of relative contribution from different pathways to the total exposure have found that food is of more importance for PFOS compared to PFOA (Gebbink et al. 2015a). Other pathways such as air and dust were of more significance for PFCAs than for PFOS, for example the contribution from precursor compounds in dust was the dominant pathway for PFOA and PFNA in a high exposure sce- nario.

Exposure and elimination pathways differ among men, women, and children. During pregnancy, PFASs are transferred through the placenta to the fetus, which is a sink for the mother but a source for the fetus (Manzano- Salgado et al. 2015). After giving birth, PFASs continues to be transferred through lactation (Karrman et al. 2007). Women eliminate PFASs through menstruation (Wong et al. 2014). Children spend more time on the floor and are therefore more exposed to dust. Therefore, when conducting exposure assessment, it is important to consider all these factors.

1. 6. Toxicity

PFCAs and PFSAs are extremely persistent and are well absorbed in the gastrointestinal tract after ingestion and are distributed mainly to the serum, liver and kidney, where they bind to serum albumin and fatty acid proteins. Persistent PFASs are not metabolized, they are excreted mainly through urine and to a lesser extent in feces. The adverse effects of PFASs include immunotoxicity, developmental toxicity, neurotoxicity, hepatotox- icity, tumor induction, weight loss, and endocrine disruption (DeWitt 2015).

Epidemiological studies have linked some PFASs to kidney and testicular cancer (Barry et al. 2013), low birth weight (Darrow et al. 2013), immune dysfunction (Grandjean et al. 2016a), thyroid disease (Melzer et al. 2010), reduced fertility for women (Fei et al. 2009), early menopause (Taylor et al. 2014), and increased cholesterol levels (Nelson et al. 2010).

(29)

The toxicity varies among PFASs depending on their structure, such as degree of fluorination, chain length, and active head group. Most toxico- logical studies have focused on PFOS and PFOA.

The risk of precursor compounds though are not only their potential to degrade into persistent and documented toxic PFASs, but also the harmful effects the precursor themselves may cause, and additionally the potential harmful effects from intermediates formed during degradation. Limited data on toxicity are available for precursor compounds. One study has shown endocrine disruption potential for PAPs in terms of inhibited male sex hormone synthesis (Rosenmai et al. 2013). In a H295R steroidogenesis assay, 8:2 monoPAP and 8:2 diPAP decreased levels of testosterone, dehy- droepiandrosterone, androstenedione, and increased levels of estrone.

Additionally, aromatase mRNA expression increased with 8:2 monoPAP and 8:2 diPAP, which could be a contribution factor to increased estrogen and decreased androgen levels.

PFPAs have shown to induce changes in apolipoprotein A-IV, related to the fatty acid metabolism, with greater effect than those of PFOS and PFOA in rat hepatoma cells (Jones et al. 2010). PFDPA was the most po- tent of three PFPAs tested.

The potential toxicity of intermediates and metabolites formed during degradation of precursor compounds have shown to have a reversed rela- tionship with chain length, compared to PFCAs and PFSAs. While in general toxicity increase with chain length, FTALs and FTUALs have been reported to have enhanced toxicity for homologues with shorter chain length, where 6:2 FTAL was more toxic for human liver cells than 8:2 FTAL, and 6:2 FTUAL was more toxic than 8:2 FTUAL (Rand et al.

2014).

1. 7. Temporal trends in humans

The use of temporal and spatial monitoring of PFASs in humans and environment are an important tool to elucidate the effects of regulations and changes in production and consumption. After the early report of PFOS in general population, increased attention and concern led to several biomonitoring studies of PFASs (Hansen et al. 2001). Retrospective studies have shown that PFAS levels increased globally in humans since the 1970-80s

(30)

until year 2000 (Haug et al. 2009, Sundstrom et al. 2011, Schroter- Kermani et al. 2013, Yeung et al. 2013b, Yeung et al. 2013c).

After year 2000 the trends differ among homologues, and spatial variations can be observed for most PFAS homologues. For PFOS, the levels started to decrease around 2000 in North America, Norway, Sweden, and Australia, which could be linked to the phase-out by the 3M Company (Calafat et al. 2007, Olsen et al. 2008, Haug et al. 2009, Glynn et al.

2012). For instance, the PFOS median level in serum decreased from 30 ng/g to 21 ng/g between 1999/2000 and 2003/2004 in the US (Calafat et al. 2007). In contrast, no change in PFOS levels have been observed in Korea between 1994 and 2008 (Harada et al. 2010). PFOS precursors have been observed in human serum and generally seem to follow the same declining trend as PFOS (Yeung et al. 2013c, Gebbink et al. 2015b).

The levels of PFOA have decreased globally since 2000 at a slower rate than PFOS, whereas an increase has been observed for long-chain (C9 – C11) PFCAs (Calafat et al. 2007, Olsen et al. 2008, Kato et al. 2011, Nost et al. 2014, De Felip et al. 2015). The trend may be attributed to the increased production of fluorotelomer compounds and contribution from fluorotelomer-based precursors. A shift in trends have been noted around 2006. Some studies reported a peak in PFNA concentrations at this time point, and also the increasing trends of PFDA and PFUnDA seem to slow down (Nost et al. 2014, Toms et al. 2014). This could be related to the initiation of the Stewardship program.

The PFSA and PFCA profile in human serum around year 2000 was generally dominated by PFOS, followed by PFOA>PFHxS>PFNA (Nost et al. 2014, Toms et al. 2014, Gribble et al. 2015). This profile has gradually altered, and long-chain PFCAs (>C8) are becoming relatively more important.

Analysis of total fluorine (TF) and extractable organic fluorine (EOF) has revealed that a large proportion of PFAS in human serum cannot be explained by the PFAS compounds generally included in biomonitoring studies (Yeung et al. 2008). A proportion of only 30 – 70% of TF could be explained by PFCAs, PFSAs, and PFOSA in Chinese serum samples from 2004. Precursor compounds could be part of the unknown PFASs. In 2009 diPAPs were detected for the first time in human serum samples from the US (D'eon et al. 2009). More recently, human serum from Germany, Chi-

(31)

na, and Sweden have been analyzed for diPAPs (Yeung et al. 2013b, Gebbink et al. 2015b, Yeung and Mabury 2016). While diPAPs were detected in European samples, no diPAPs were found in the Chinese samples, which indicates geographical differences in PFAS exposure patterns. In German samples, temporal trends between 1982 and 2009 were assessed, revealing increasing levels of long-chain PFCAs and decreasing trend for PFOA in recent years (2000 – 2009), and no change in concentration for diPAPs. Another emergent PFAS class, PFPiA, was detected in 2011 in human serum (Lee and Mabury 2011).

The FTSAs were detected for the first time in human serum in 2002, and have later been detected in serum from US, Germany, and China (Connol- ly et al. 2002, Lee and Mabury 2011, Yeung and Mabury 2016).

In general, low PFCA precursor levels in human serum have been reported. That does not rule out the possibility that precursors are of significance for human exposure, since they may have been readily biodegraded to corresponding PFCAs.

1. 8. Temporal trend in the environment

Concerns about the impact of PFASs to the environment arose in 2001, when global PFOS contamination in wild-life was reported for the first time (Giesy and Kannan 2001). Environmental contamination can be tracked backed to the 1950s. Analysis of sediment cores from Canada has shown that PFOS can be observed with beginning in 1952, whereas FOSA and long-chain PFCAs (C8 – C10) appear first in 1970s. PFAS concentrations continues to increase during the whole study period until 2005 (Yeung et al. 2013b). The patterns observed in the sediment cores are in temporal agreement with appearance of new compounds entering the PFAS market and industry, but are not reflecting the turning point for ceased production of PFOS by the 3M Company.

In wildlife, a number of species of fish, birds, and mammals have been studied, especially in the aquatic environment and in the northern hemi- sphere. Various PFAS trends have been reported, but a common trend is a significant increase of PFOS since 1970s up to around 2000 (Paul et al.

2009). In the marine environment, increasing trends in polar bears, guil- lemots, and pilot whales were reported (Smithwick et al. 2006, Holmstrom et al. 2010, Rotander et al. 2012). Geographical variations

(32)

where also observed within the same specie; as for herring gull where levels increased in the Baltic Sea but not in the North Sea (Rudel et al. 2011);

and for ringed seals where levels increased between 1994 and 2003 in East Greenland, but not in West Greenland (Bossi et al. 2005). In the terrestrial environment, increasing trend has been less apparent compared to the marine environment; for instance in roe deer where levels slightly increased between 1989 and 2001 (Falk et al. 2012).

After the 3M Company’s phase-out, there are various trends of PFOS in the environment. For some species and locations, PFOS has been reported to decrease after 2000, as in in roe deer (Falk et al. 2012), and sea otter (Hart 2009 temporal trends). A peak around 2000 has been observed for ringed seals and polar bears (Riget et al. 2013). In other cases, little change in PFOS levels are observed, for example in seals, dolphins, and whales, and in golden eagles from the terrestrial environment (Rotander et al. 2012, Herzke et al. 2014). Increasing levels has also been reported, for example in otters from Sweden the PFOS levels continuously increased from 1972 to 2011 (Roos et al. 2013).

The trends of PFCAs are more complicated than that of PFOS. Levels increased between 2000 and 2006 – 2008; after that levels of PFOA and PFNA declined, while other long-chain PFCA levels either stabilized or increased. Increased levels of long-chain PFCAs since 2000 have been observed in the marine environment, for example in seals, polar bears, har- bor porpoise, and whales (Routti et al. 2011, Huber et al. 2012, Rotander et al. 2012, Riget et al. 2013). In the terrestrial environment, increasing levels of long-chain PFCAs are also observed, as for tawny owl from Norway and otters from Sweden (Ahrens et al. 2011b, Roos et al. 2013).

On the other hand, no apparent trend was observed for PFNA and PFDA in European roe deer after 2001 (Falk et al. 2012).

Only a few studies have reported PFCA precursors in wildlife, therefore temporal trends for these compound classes are lacking. However, PAPs and FTSAs have been found globally in the environment. In the last dec- ade, PAPs have been observed in lake trout from Canada sampled in 2009 (Guo et al. 2012), in mussels from Spain in 2009, in benthic worms from HongKong in 2011 (Loi et al. 2013), in zooplankton from Baltic Sea (Gebbink et al. 2016a), and in tuna from Indian Ocean in 2013 (Zabaleta et al. 2015). On the other hand, PAPs were not detected in fish from Baltic

(33)

Sea. The 6:2 FTSA has been found in ice amphipod sampled in 2004 (Haukas et al. 2007), in benthic worms from HongKong in 2011 (Loi et al. 2013), and in fish from a contaminated area in Norway in 2011 (Karrman et al. 2011).

Global measurements of FTOHs in the atmosphere between 2006 and 2011 have shown increasing levels of 6:2 FTOH (Gawor et al. 2014).

While 10:2 FTOH decreased during these studied periods, 8:2 FTOH ini- tially declined but returned to initial level in 2011. Levels of MeFOSA and MeFOSE decreased between 2006 and 2011.

Besides detection of PFCA precursors, product change towards short-chain PFASs is also reflected by recent reports on PFBS and PFBS precursors in the environment. Levels of PFBS in cetacean have increased in the South China Sea between 2002 and 2014 (Lam et al. 2016). PFBS has been found in mammals in Greenland (Gebbink et al. 2016b). Fish from Cana- da and Europe has been found to have levels of perfluorobutane sulfonamide (FBSA) at up to 80 ng/g (Chu et al. 2016). In addition, methyl perfluorobutane sulfonamidoethanol (MeFBSE) and methyl perfluorobutane sulfonamide (MeFBSA) have been detected in the atmosphere over the North China Sea, at comparable levels as PFOS precursors (Lai et al.

2016).

To summarize, a scarce number of studies reports about emerging PFASs, including precursors, and observations made so far witness of global oc- currence and distribution in various types of environmental compartments.

(34)

2. Aim and objectives

The aim of the thesis was to assess if polyfluoroalkyl precursor compounds are a significant contributor to PFCA exposure, both environmental and human exposure. Such exposure data of other compounds besides persistent perfluoroalkyl substances is an important basis for risk assessments and not least for regulatory and policy work aiming at reducing hazardous chemicals in the society. The hypothesis of the work was that current monitoring and regulatory efforts are insufficient to protect humans and the environment from PFASs. By a cross-section analysis of the biosphere and technosphere the contribution of precursor compounds in relation to persistent PFASs in humans and the environment can be assessed. The specific objectives were:

• Assess human levels of semi-persistent precursor compounds in relation to persistent PFAS by analysis of sera

• Assess human exposure of semi-persistent precursor compounds in relation to persistent PFAS from indoor environment by analysis of household dust

• Assess the release of semi-persistent precursor compounds in relation to persistent PFAS from households to the environment by analysis of waste water and sewage sludge

• Assess the environmental exposure of semi-persistent precursor compounds in relation to persistent PFAS by analysis of wild bird eggs

Contribution of polyfluoroalkyl phosphate esters (PAPs) and other precursor compounds to perfluoroalkyl carboxylates (PFCAs) in humans and the environment

Contribution of polyfluoroalkyl phosphate esters (PAPs) and other precursor compounds to perfluoroalkyl carboxylates (PFCAs) in humans and the environment

Abstract

Abstract

Abstract

List of papers

List of abbreviations

Table of Contents

1. Introduction

2. Aim and objectives